soe-filter

To filter state-owned stocks from an A-share investment universe, apply exclusion rules to identify and drop central-enterprise equities. The skill processes stock data and outputs a non-SOE universe parquet alongside a review CSV for ambiguous cases.

Yes, state-owned enterprise screening accepts a DuckDB stocks table as input for filtering. It also supports reading directly from a Parquet file and gracefully falls back to a cache CSV when primary data sources are unavailable.

The R1-R4 exclusion rules are four applied rule sets designed to identify and drop state-owned and central-enterprise stocks from equity universes. They ensure comprehensive filtering while isolating uncertain cases into a separate review CSV.

During SOE universe filtering, ambiguous stocks that cannot be definitively classified by the exclusion rules are separated. The skill writes these uncertain cases to a review CSV, allowing manual inspection while retaining clear non-SOE stocks in a parquet file.

Yes, stock screening for state-owned enterprises works with pandas and Parquet files. The skill leverages both pandas and DuckDB dependencies to process equity data and writes the resulting non-SOE universe directly to a Parquet file.

If an A-share universe retains state-owned equities after filtering, they are likely ambiguous cases flagged by the four exclusion rules. Check the generated review CSV, which captures uncertain stocks that require manual verification before removal.